1 /*****************************************************************************/
5 /* Source file line info structure */
9 /* (C) 1998-2010, Ullrich von Bassewitz */
10 /* Roemerstrasse 52 */
11 /* D-70794 Filderstadt */
12 /* EMail: uz@cc65.org */
15 /* This software is provided 'as-is', without any expressed or implied */
16 /* warranty. In no event will the authors be held liable for any damages */
17 /* arising from the use of this software. */
19 /* Permission is granted to anyone to use this software for any purpose, */
20 /* including commercial applications, and to alter it and redistribute it */
21 /* freely, subject to the following restrictions: */
23 /* 1. The origin of this software must not be misrepresented; you must not */
24 /* claim that you wrote the original software. If you use this software */
25 /* in a product, an acknowledgment in the product documentation would be */
26 /* appreciated but is not required. */
27 /* 2. Altered source versions must be plainly marked as such, and must not */
28 /* be misrepresented as being the original software. */
29 /* 3. This notice may not be removed or altered from any source */
32 /*****************************************************************************/
64 /*****************************************************************************/
66 /*****************************************************************************/
70 Token CurTok; /* The current token */
71 Token NextTok; /* The next token */
77 TT_C89 = 0x01 << STD_C89, /* Token valid in C89 */
78 TT_C99 = 0x01 << STD_C99, /* Token valid in C99 */
79 TT_CC65 = 0x01 << STD_CC65 /* Token valid in cc65 */
83 static const struct Keyword {
84 char* Key; /* Keyword name */
85 unsigned char Tok; /* The token */
86 unsigned char Std; /* Token supported in which standards? */
88 { "_Pragma", TOK_PRAGMA, TT_C89 | TT_C99 | TT_CC65 }, /* !! */
89 { "__AX__", TOK_AX, TT_C89 | TT_C99 | TT_CC65 },
90 { "__A__", TOK_A, TT_C89 | TT_C99 | TT_CC65 },
91 { "__EAX__", TOK_EAX, TT_C89 | TT_C99 | TT_CC65 },
92 { "__X__", TOK_X, TT_C89 | TT_C99 | TT_CC65 },
93 { "__Y__", TOK_Y, TT_C89 | TT_C99 | TT_CC65 },
94 { "__asm__", TOK_ASM, TT_C89 | TT_C99 | TT_CC65 },
95 { "__attribute__", TOK_ATTRIBUTE, TT_C89 | TT_C99 | TT_CC65 },
96 { "__cdecl__", TOK_CDECL, TT_C89 | TT_C99 | TT_CC65 },
97 { "__far__", TOK_FAR, TT_C89 | TT_C99 | TT_CC65 },
98 { "__fastcall__", TOK_FASTCALL, TT_C89 | TT_C99 | TT_CC65 },
99 { "__inline__", TOK_INLINE, TT_C89 | TT_C99 | TT_CC65 },
100 { "__near__", TOK_NEAR, TT_C89 | TT_C99 | TT_CC65 },
101 { "asm", TOK_ASM, TT_CC65 },
102 { "auto", TOK_AUTO, TT_C89 | TT_C99 | TT_CC65 },
103 { "break", TOK_BREAK, TT_C89 | TT_C99 | TT_CC65 },
104 { "case", TOK_CASE, TT_C89 | TT_C99 | TT_CC65 },
105 { "cdecl", TOK_CDECL, TT_CC65 },
106 { "char", TOK_CHAR, TT_C89 | TT_C99 | TT_CC65 },
107 { "const", TOK_CONST, TT_C89 | TT_C99 | TT_CC65 },
108 { "continue", TOK_CONTINUE, TT_C89 | TT_C99 | TT_CC65 },
109 { "default", TOK_DEFAULT, TT_C89 | TT_C99 | TT_CC65 },
110 { "do", TOK_DO, TT_C89 | TT_C99 | TT_CC65 },
111 { "double", TOK_DOUBLE, TT_C89 | TT_C99 | TT_CC65 },
112 { "else", TOK_ELSE, TT_C89 | TT_C99 | TT_CC65 },
113 { "enum", TOK_ENUM, TT_C89 | TT_C99 | TT_CC65 },
114 { "extern", TOK_EXTERN, TT_C89 | TT_C99 | TT_CC65 },
115 { "far", TOK_FAR, TT_CC65 },
116 { "fastcall", TOK_FASTCALL, TT_CC65 },
117 { "float", TOK_FLOAT, TT_C89 | TT_C99 | TT_CC65 },
118 { "for", TOK_FOR, TT_C89 | TT_C99 | TT_CC65 },
119 { "goto", TOK_GOTO, TT_C89 | TT_C99 | TT_CC65 },
120 { "if", TOK_IF, TT_C89 | TT_C99 | TT_CC65 },
121 { "inline", TOK_INLINE, TT_C99 | TT_CC65 },
122 { "int", TOK_INT, TT_C89 | TT_C99 | TT_CC65 },
123 { "long", TOK_LONG, TT_C89 | TT_C99 | TT_CC65 },
124 { "near", TOK_NEAR, TT_CC65 },
125 { "register", TOK_REGISTER, TT_C89 | TT_C99 | TT_CC65 },
126 { "restrict", TOK_RESTRICT, TT_C99 | TT_CC65 },
127 { "return", TOK_RETURN, TT_C89 | TT_C99 | TT_CC65 },
128 { "short", TOK_SHORT, TT_C89 | TT_C99 | TT_CC65 },
129 { "signed", TOK_SIGNED, TT_C89 | TT_C99 | TT_CC65 },
130 { "sizeof", TOK_SIZEOF, TT_C89 | TT_C99 | TT_CC65 },
131 { "static", TOK_STATIC, TT_C89 | TT_C99 | TT_CC65 },
132 { "struct", TOK_STRUCT, TT_C89 | TT_C99 | TT_CC65 },
133 { "switch", TOK_SWITCH, TT_C89 | TT_C99 | TT_CC65 },
134 { "typedef", TOK_TYPEDEF, TT_C89 | TT_C99 | TT_CC65 },
135 { "union", TOK_UNION, TT_C89 | TT_C99 | TT_CC65 },
136 { "unsigned", TOK_UNSIGNED, TT_C89 | TT_C99 | TT_CC65 },
137 { "void", TOK_VOID, TT_C89 | TT_C99 | TT_CC65 },
138 { "volatile", TOK_VOLATILE, TT_C89 | TT_C99 | TT_CC65 },
139 { "while", TOK_WHILE, TT_C89 | TT_C99 | TT_CC65 },
141 #define KEY_COUNT (sizeof (Keywords) / sizeof (Keywords [0]))
145 /* Stuff for determining the type of an integer constant */
149 #define IT_ULONG 0x08
153 /*****************************************************************************/
155 /*****************************************************************************/
159 static int CmpKey (const void* Key, const void* Elem)
160 /* Compare function for bsearch */
162 return strcmp ((const char*) Key, ((const struct Keyword*) Elem)->Key);
167 static token_t FindKey (const char* Key)
168 /* Find a keyword and return the token. Return IDENT if the token is not a
173 K = bsearch (Key, Keywords, KEY_COUNT, sizeof (Keywords [0]), CmpKey);
174 if (K && (K->Std & (0x01 << IS_Get (&Standard))) != 0) {
183 static int SkipWhite (void)
184 /* Skip white space in the input stream, reading and preprocessing new lines
185 ** if necessary. Return 0 if end of file is reached, return 1 otherwise.
189 while (CurC == '\0') {
190 if (NextLine () == 0) {
195 if (IsSpace (CurC)) {
205 int TokIsFuncSpec (const Token* T)
206 /* Return true if the token is a function specifier */
208 return (T->Tok == TOK_INLINE) ||
209 (T->Tok == TOK_FASTCALL) || (T->Tok == TOK_CDECL) ||
210 (T->Tok == TOK_NEAR) || (T->Tok == TOK_FAR);
215 void SymName (char* S)
216 /* Read a symbol from the input stream. The first character must have been
217 ** checked before calling this function. The buffer is expected to be at
218 ** least of size MAX_IDENTLEN+1.
223 if (Len < MAX_IDENTLEN) {
228 } while (IsIdent (CurC) || IsDigit (CurC));
235 /* If a symbol follows, read it and return 1, otherwise return 0 */
237 if (IsIdent (CurC)) {
247 static void UnknownChar (char C)
248 /* Error message for unknown character */
250 Error ("Invalid input character with code %02X", C & 0xFF);
251 NextChar (); /* Skip */
256 static void SetTok (int tok)
257 /* Set NextTok.Tok and bump line ptr */
265 static int ParseChar (void)
266 /* Parse a character. Converts escape chars into character codes. */
272 /* Check for escape chars */
311 /* Hex character constant */
312 if (!IsXDigit (NextC)) {
313 Error ("\\x used with no following hex digits");
318 while (IsXDigit (NextC)) {
319 if ((C << 4) >= 256) {
321 Error ("Hex character constant out of range");
325 C = (C << 4) | HexVal (NextC);
342 while (IsODigit (NextC) && Count++ < 3) {
343 C = (C << 3) | HexVal (NextC);
347 Error ("Octal character constant out of range");
350 Error ("Illegal character constant");
352 /* Try to do error recovery, otherwise the compiler will spit
353 ** out thousands of errors in this place and abort.
355 if (CurC != '\'' && CurC != '\0') {
356 while (NextC != '\'' && NextC != '\"' && NextC != '\0') {
366 /* Skip the character read */
369 /* Do correct sign extension */
370 return SignExtendChar (C);
375 static void CharConst (void)
376 /* Parse a character constant. */
386 /* Check for closing quote */
388 Error ("`\'' expected");
394 /* Setup values and attributes */
395 NextTok.Tok = TOK_CCONST;
397 /* Translate into target charset */
398 NextTok.IVal = SignExtendChar (TgtTranslateChar (C));
400 /* Character constants have type int */
401 NextTok.Type = type_int;
406 static void StringConst (void)
407 /* Parse a quoted string */
410 StrBuf S = AUTO_STRBUF_INITIALIZER;
412 /* Assume next token is a string constant */
413 NextTok.Tok = TOK_SCONST;
415 /* Concatenate strings. If at least one of the concenated strings is a wide
416 ** character literal, the whole string is a wide char literal, otherwise
417 ** it's a normal string literal.
421 /* Check if this is a normal or a wide char string */
422 if (CurC == 'L' && NextC == '\"') {
423 /* Wide character literal */
424 NextTok.Tok = TOK_WCSCONST;
427 } else if (CurC == '\"') {
428 /* Skip the quote char */
435 /* Read until end of string */
436 while (CurC != '\"') {
438 Error ("Unexpected newline");
441 SB_AppendChar (&S, ParseChar ());
444 /* Skip closing quote char if there was one */
447 /* Skip white space, read new input */
452 /* Terminate the string */
453 SB_AppendChar (&S, '\0');
455 /* Add the whole string to the literal pool */
456 NextTok.SVal = AddLiteralStr (&S);
458 /* Free the buffer */
464 static void NumericConst (void)
465 /* Parse a numeric constant */
467 unsigned Base; /* Temporary number base */
468 unsigned Prefix; /* Base according to prefix */
469 StrBuf S = STATIC_STRBUF_INITIALIZER;
473 unsigned long IVal; /* Value */
475 /* Check for a leading hex or octal prefix and determine the possible
479 /* Gobble 0 and examine next char */
481 if (toupper (CurC) == 'X') {
483 NextChar (); /* gobble "x" */
485 Base = 10; /* Assume 10 for now - see below */
486 Prefix = 8; /* Actual prefix says octal */
492 /* Because floating point numbers don't have octal prefixes (a number
493 ** with a leading zero is decimal), we first have to read the number
494 ** before converting it, so we can determine if it's a float or an
497 while (IsXDigit (CurC) && HexVal (CurC) < Base) {
498 SB_AppendChar (&S, CurC);
503 /* The following character tells us if we have an integer or floating
504 ** point constant. Note: Hexadecimal floating point constants aren't
507 IsFloat = (CurC == '.' ||
508 (Base == 10 && toupper (CurC) == 'E') ||
509 (Base == 16 && toupper (CurC) == 'P' && IS_Get (&Standard) >= STD_C99));
511 /* If we don't have a floating point type, an octal prefix results in an
514 if (!IsFloat && Prefix == 8) {
518 /* Since we do now know the correct base, convert the remembered input
523 while ((C = SB_Get (&S)) != '\0') {
524 DigitVal = HexVal (C);
525 if (DigitVal >= Base) {
526 Error ("Numeric constant contains digits beyond the radix");
528 IVal = (IVal * Base) + DigitVal;
531 /* We don't need the string buffer any longer */
534 /* Distinguish between integer and floating point constants */
540 /* Check for a suffix and determine the possible types */
542 if (toupper (CurC) == 'U') {
545 if (toupper (CurC) != 'L') {
546 Types = IT_UINT | IT_ULONG;
551 } else if (toupper (CurC) == 'L') {
554 if (toupper (CurC) != 'U') {
555 Types = IT_LONG | IT_ULONG;
563 /* Decimal constants are of any type but uint */
564 Types = IT_INT | IT_LONG | IT_ULONG;
566 /* Octal or hex constants are of any type */
567 Types = IT_INT | IT_UINT | IT_LONG | IT_ULONG;
571 /* Check the range to determine the type */
573 /* Out of range for int */
575 /* If the value is in the range 0x8000..0xFFFF, unsigned int is not
576 ** allowed, and we don't have a type specifying suffix, emit a
577 ** warning, because the constant is of type long.
579 if (IVal <= 0xFFFF && (Types & IT_UINT) == 0 && !HaveSuffix) {
580 Warning ("Constant is long");
584 /* Out of range for unsigned int */
587 if (IVal > 0x7FFFFFFF) {
588 /* Out of range for long int */
592 /* Now set the type string to the smallest type in types */
593 if (Types & IT_INT) {
594 NextTok.Type = type_int;
595 } else if (Types & IT_UINT) {
596 NextTok.Type = type_uint;
597 } else if (Types & IT_LONG) {
598 NextTok.Type = type_long;
600 NextTok.Type = type_ulong;
603 /* Set the value and the token */
605 NextTok.Tok = TOK_ICONST;
610 Double FVal = FP_D_FromInt (IVal); /* Convert to double */
612 /* Check for a fractional part and read it */
620 /* Read fractional digits */
621 Scale = FP_D_Make (1.0);
622 while (IsXDigit (CurC) && (DigitVal = HexVal (CurC)) < Base) {
623 /* Get the value of this digit */
624 Double FracVal = FP_D_Div (FP_D_FromInt (DigitVal * Base), Scale);
625 /* Add it to the float value */
626 FVal = FP_D_Add (FVal, FracVal);
628 Scale = FP_D_Mul (Scale, FP_D_FromInt (DigitVal));
634 /* Check for an exponent and read it */
635 if ((Base == 16 && toupper (CurC) == 'F') ||
636 (Base == 10 && toupper (CurC) == 'E')) {
641 /* Skip the exponent notifier */
644 /* Read an optional sign */
647 } else if (CurC == '+') {
651 /* Read exponent digits. Since we support only 32 bit floats
652 ** with a maximum exponent of +-/127, we read the exponent
653 ** part as integer with up to 3 digits and drop the remainder.
654 ** This avoids an overflow of Exp. The exponent is always
655 ** decimal, even for hex float consts.
659 while (IsDigit (CurC)) {
661 Exp = Exp * 10 + HexVal (CurC);
666 /* Check for errors: We must have exponent digits, and not more
670 Error ("Floating constant exponent has no digits");
671 } else if (Digits > 3) {
672 Warning ("Floating constant exponent is too large");
675 /* Scale the exponent and adjust the value accordingly */
677 FVal = FP_D_Mul (FVal, FP_D_Make (pow (10, Exp)));
681 /* Check for a suffix and determine the type of the constant */
682 if (toupper (CurC) == 'F') {
684 NextTok.Type = type_float;
686 NextTok.Type = type_double;
689 /* Set the value and the token */
691 NextTok.Tok = TOK_FCONST;
698 void NextToken (void)
699 /* Get next token from input stream */
703 /* We have to skip white space here before shifting tokens, since the
704 ** tokens and the current line info is invalid at startup and will get
705 ** initialized by reading the first time from the file. Remember if
706 ** we were at end of input and handle that later.
708 int GotEOF = (SkipWhite() == 0);
710 /* Current token is the lookahead token */
712 ReleaseLineInfo (CurTok.LI);
716 /* When reading the first time from the file, the line info in NextTok,
717 ** which was copied to CurTok is invalid. Since the information from
718 ** the token is used for error messages, we must make it valid.
720 if (CurTok.LI == 0) {
721 CurTok.LI = UseLineInfo (GetCurLineInfo ());
724 /* Remember the starting position of the next token */
725 NextTok.LI = UseLineInfo (GetCurLineInfo ());
727 /* Now handle end of input. */
729 /* End of file reached */
730 NextTok.Tok = TOK_CEOF;
734 /* Determine the next token from the lookahead */
735 if (IsDigit (CurC) || (CurC == '.' && IsDigit (NextC))) {
741 /* Check for wide character literals */
742 if (CurC == 'L' && NextC == '\"') {
747 /* Check for keywords and identifiers */
750 /* Check for a keyword */
751 if ((NextTok.Tok = FindKey (token)) != TOK_IDENT) {
752 /* Reserved word found */
755 /* No reserved word, check for special symbols */
756 if (token[0] == '_' && token[1] == '_') {
757 /* Special symbols */
758 if (strcmp (token+2, "FILE__") == 0) {
759 NextTok.SVal = AddLiteral (GetCurrentFile());
760 NextTok.Tok = TOK_SCONST;
762 } else if (strcmp (token+2, "LINE__") == 0) {
763 NextTok.Tok = TOK_ICONST;
764 NextTok.IVal = GetCurrentLine();
765 NextTok.Type = type_int;
767 } else if (strcmp (token+2, "func__") == 0) {
768 /* __func__ is only defined in functions */
770 NextTok.SVal = AddLiteral (F_GetFuncName (CurrentFunc));
771 NextTok.Tok = TOK_SCONST;
777 /* No reserved word but identifier */
778 strcpy (NextTok.Ident, token);
779 NextTok.Tok = TOK_IDENT;
783 /* Monstrous switch statement ahead... */
791 NextTok.Tok = TOK_BOOL_NOT;
802 SetTok (TOK_MOD_ASSIGN);
804 NextTok.Tok = TOK_MOD;
812 SetTok (TOK_BOOL_AND);
815 SetTok (TOK_AND_ASSIGN);
818 NextTok.Tok = TOK_AND;
837 SetTok (TOK_MUL_ASSIGN);
839 NextTok.Tok = TOK_STAR;
850 SetTok (TOK_PLUS_ASSIGN);
853 NextTok.Tok = TOK_PLUS;
868 SetTok (TOK_MINUS_ASSIGN);
871 SetTok (TOK_PTR_REF);
874 NextTok.Tok = TOK_MINUS;
883 SetTok (TOK_ELLIPSIS);
888 NextTok.Tok = TOK_DOT;
895 SetTok (TOK_DIV_ASSIGN);
897 NextTok.Tok = TOK_DIV;
918 SetTok (TOK_SHL_ASSIGN);
920 NextTok.Tok = TOK_SHL;
924 NextTok.Tok = TOK_LT;
933 NextTok.Tok = TOK_ASSIGN;
946 SetTok (TOK_SHR_ASSIGN);
948 NextTok.Tok = TOK_SHR;
952 NextTok.Tok = TOK_GT;
971 SetTok (TOK_XOR_ASSIGN);
973 NextTok.Tok = TOK_XOR;
985 SetTok (TOK_BOOL_OR);
988 SetTok (TOK_OR_ASSIGN);
991 NextTok.Tok = TOK_OR;
1012 void SkipTokens (const token_t* TokenList, unsigned TokenCount)
1013 /* Skip tokens until we reach TOK_CEOF or a token in the given token list.
1014 ** This routine is used for error recovery.
1017 while (CurTok.Tok != TOK_CEOF) {
1019 /* Check if the current token is in the token list */
1021 for (I = 0; I < TokenCount; ++I) {
1022 if (CurTok.Tok == TokenList[I]) {
1023 /* Found a token in the list */
1028 /* Not in the list: Skip it */
1036 int Consume (token_t Token, const char* ErrorMsg)
1037 /* Eat token if it is the next in the input stream, otherwise print an error
1038 ** message. Returns true if the token was found and false otherwise.
1041 if (CurTok.Tok == Token) {
1045 Error ("%s", ErrorMsg);
1052 int ConsumeColon (void)
1053 /* Check for a colon and skip it. */
1055 return Consume (TOK_COLON, "`:' expected");
1060 int ConsumeSemi (void)
1061 /* Check for a semicolon and skip it. */
1063 /* Try do be smart about typos... */
1064 if (CurTok.Tok == TOK_SEMI) {
1068 Error ("`;' expected");
1069 if (CurTok.Tok == TOK_COLON || CurTok.Tok == TOK_COMMA) {
1078 int ConsumeComma (void)
1079 /* Check for a comma and skip it. */
1081 /* Try do be smart about typos... */
1082 if (CurTok.Tok == TOK_COMMA) {
1086 Error ("`,' expected");
1087 if (CurTok.Tok == TOK_SEMI) {
1096 int ConsumeLParen (void)
1097 /* Check for a left parenthesis and skip it */
1099 return Consume (TOK_LPAREN, "`(' expected");
1104 int ConsumeRParen (void)
1105 /* Check for a right parenthesis and skip it */
1107 return Consume (TOK_RPAREN, "`)' expected");
1112 int ConsumeLBrack (void)
1113 /* Check for a left bracket and skip it */
1115 return Consume (TOK_LBRACK, "`[' expected");
1120 int ConsumeRBrack (void)
1121 /* Check for a right bracket and skip it */
1123 return Consume (TOK_RBRACK, "`]' expected");
1128 int ConsumeLCurly (void)
1129 /* Check for a left curly brace and skip it */
1131 return Consume (TOK_LCURLY, "`{' expected");
1136 int ConsumeRCurly (void)
1137 /* Check for a right curly brace and skip it */
1139 return Consume (TOK_RCURLY, "`}' expected");